Microsoft defines Azure instance lifecycle but leaves out timing
Microsoft has documented how Azure virtual machines reach the end of their life, but the new guidance still doesn't tell customers when to expect it.
Compute & Hardware
Chips, datacenters, interconnect, and the power they draw.
Alienware RTX 5090 PC Costs Less Than GPU Plus CPU Alone
A pre-built Alienware Area-51 with an RTX 5090 and Ryzen 9 9950X3D is priced below the combined street cost of those two components alone.
Olight imini 2 keychain flashlight drops to $11.99 at Amazon
The tiny rechargeable light is 40% off, making it a cheap hands-free option for PC builders and anyone working in dark spaces.
A $65 Mechanical Keyboard Surprises With Premium Switch Feel
A Tom's Hardware writer found that the budget Epomaker x Aula F75 Pro delivers switch quality that rivals far more expensive boards.
Geekom Fall Sale Offers Up to $730 Off Mini PCs
Geekom's fall sale discounts 13 mini PCs on Amazon and its own site, with the IT13 Max seeing the deepest cut.
Nvidia reportedly ends RTX 5090 production to focus on AI GPUs
Unverified reports say Nvidia has stopped making GB202 chips for GeForce RTX 5090 cards, shifting supply to higher-margin data center and professional GPUs.
Corsair 32GB DDR5-5600 Kit Drops to $414.99, Cheapest at Its Speed
A $103 discount makes Corsair's 32GB DDR5-5600 Vengeance kit the lowest-priced option at that speed, though RAM prices remain high.
AI Chip Photonics Face Heat and Stress Hurdles
Experts say co-packaged optics cannot scale until engineers model transient heat and mechanical stress across multi-die systems.
COSMIC bans AI-written code while GNOME weighs AI bug reports
System76's COSMIC desktop requires human-written contributions, while a GNOME developer argues AI-discovered flaws are too valuable to ignore.
Endpoint monitoring turns meeting rooms into proactive IT assets
Cloud-based telemetry lets IT teams treat meeting rooms as managed endpoints, shifting from reactive fixes to predictive maintenance and evidence-based investment.
Chiplet Interconnect, CFETs, and LLM Latency Top Weekly EDA Blogs
A weekly roundup of semiconductor engineering blogs highlights advances in chiplet interconnects, transistor architecture, and LLM inference latency.
Broker Pleads Guilty in Super Micro AI Chip Smuggling Case
A guilty plea in the scheme to divert Nvidia AI chips to China strengthens U.S. enforcement of export controls.
Dutch tax office to leave Microsoft 365 cloud for on-premises tools
The Dutch tax authority will move email and calendars in-house by 2027 and later adopt European open source tools, citing sovereignty concerns.
3D-Printed Jet RC Plane Kingchaser Flies, Targets Mach 0.8
A mostly 3D-printed remote-controlled aircraft with a jet turbine has made its maiden flight and aims to break the world speed record.
Laptop mod with 360mm radiator cuts temps by 60°C, boosts performance 14%
A hobbyist's extreme liquid-cooling mod shows how much thermal headroom laptops sacrifice for portability.
Modder Runs DLSS 5 on Pac-Man, With Horrifying Results
A modder injected Nvidia's leaked DLSS 5 into MS-DOS Pac-Man, producing glitchy, unsettling visuals that show the upscaler's limits on flat 2D games.
Free Steam Demo's $1,000 Daily AI Bill Forces Devs to Take Bank Loan
A free AI-driven Steam demo shows how per-user inference costs can explode when player counts spike, pushing developers to seek loans and plan for local models.
AI Labs Reportedly War-Game Catastrophic Rogue AI Scenarios
Top AI labs are privately simulating major AI-caused disruptions to critical infrastructure, signaling growing concern about near-term risks.